Samantha Sacks

Senior Design Project:

Performance Assessment of Forecasting-Integrated Dynamic-Pricing Processes via Simulation Models


Senior Systems Design Project

Advisor: Professor Yossi Aviv



This paper summarizes the procedure and results of a simulation using dynamic pricing in a single item retail industry. Concepts in revenue management such as forecasting were used to execute this project. The simulation was done in the Extend programming language. This programming language interfaced with Microsoft Excel, which was used for computations.


The project was composed of two parts. The first of which was creating the simulation of a single item retail situation. This is not common in the market, however to add more items would exponentially increase the complexity of the project. The second part of the project was finding an optimal pricing strategy for the store modeled in the simulation. The simulation is a tool which generates customer data and outputs purchase information such as time of purchase and the price the item was sold at. Different pricing strategies are tested using calculations in Microsoft Excel to find the best pricing strategy for this single item retail situation.


This project utilizes operations research methods and forecasting to create a simulation of a dynamic pricing process for a retail industry. Topics addressed from the Systems Engineering curriculum include: Simulation, Optimization, Statistics, and Operations Research.



The flow chart below illustrates how the project progressed. Initially a simulation was created which generated data used in a forecasting tool. The forecasting method created an optimal pricing strategy that was then fed back into the simulation to dynamically update the price.



The main question researched was:

(i)          What is the impact of various forecasting techniques on expected revenues?





A simulation was created in order to accurately represent the behavior of a retail situation. In the simulation customers are generated according to a Poisson distribution. The customers are randomly assigned a reservation price, or the price which they are willing to pay for the good in the simulation. As each customer passes through the service node of the simulation, their reservation price is compared to the good’s listed price and either a purchase is made or the customer leaves the node and incurs a time delay, which represents a customer waiting for the price of the good to drop. The system was designed using the Extend simulation software, which interacts with Microsoft Excel.


In Excel, data is recorded and continuously updated as the state of the system changes. The data collected in Excel is input to a forecasting model. From this forecasting model a continuously changing price is put back into the simulation model in Extend.


Summary of Results:

Of the three pricing strategies tested the one which yielded the best result was a strategy based on targeting a conversion ratio. This strategy allowed the item price to change according to information output in the simulation. As well the pricing adjusted so that a majority of customers made purchases, hypothetically bringing in a great deal of revenue.

Courses and Concepts used from the Systems Science Curriculum:


~ Applied Operations Research – Extend programming language

~ Operations Research Optimization techniques

~ Engineering Probability and Statistics – probability distribution concepts




~ Talluri, Kaylan T. and Van Ryzin, Garrett J. The Theory and Practice of Revenue Management. New York: Springer Science+Business Media. 2005.